Ramucirumab plus pembrolizumab in previously treated advanced NSCLC and gastro-oesophageal and urothelial carcinomas: non-randomised, multi-cohort, open-label, Phase 1 trial (JVDF)

Professor Roy S. Herbst, MD,1* Hendrik-Tobias Arkenau, MD,2 Rafael Santana-Davila, MD,3 Emiliano Calvo, MD,4 Professor Luis Paz-Ares, MD,5 Philippe A. Cassier, MD,6 Johanna Bendell, MD,7 Professor Nicolas Penel, MD,8 Matthew G. Krebs, MD,9 Juan Martin-Liberal, MD,10,11 Nicolas Isambert, MD,12 Andres Soriano, MD,13 Martin Wermke, MD,14 Jennifer Cultrera, MD,13 Ling Gao, PhD,15 Ryan C Widau, PhD,16 Gu Mi, PhD,16 Jin Jin, PhD,16 David Ferry, MD,15 Professor Charles Fuchs, MD,1 Professor Daniel P. Petrylak, MD,1 Ian Chau, MD17

1Yale University School of Medicine, Yale Cancer Center, New Haven, CT, USA (R. Herbst, MD; C. Fuchs, MD; D. Petrylak, MD) 2Drug Development Unit, Sarah Cannon Research Institute UK, and Cancer Institute, University College London, London, UK (H-T. Arkenau, MD)3University of Washington, Seattle, WA, USA (R. Santana-Davila, MD)4Early Clinical Drug Development Program, START Madrid-HM CIOCC, Centro Integral Oncológico Clara Campal, Madrid, Spain (E. Calvo, MD)5Virgen del Rocio University Hospital, Seville, Spain (L. Paz-Ares, MD)6Centre Léon Bérard, Lyon, France (P. Cassier, MD)7Sarah Cannon Research Institute/Tennessee Oncology, Nashville, TN, USA (J. Bendell, MD8Centre Oscar Lambret and Lille University, Lille, France (N. Penel, MD)9 Division of Cancer Sciences, Faculty of Biology, Medicine and Health, University of Manchester and The Christie NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK (M. Krebs, MD)10Vall d'Hebron Institute of Oncology (VHIO), Barcelona, Spain (J. Martin-Liberal, MD)11Catalan Institute of Oncology (ICO) Hospitalet, Barcelona, Spain (J. Martin-Liberal, MD)12Centre Georges François Leclerc, Dijon, France (N. Isambert, MD)13Florida Cancer Specialists, Englewood, FL, USA (A. Soriano, MD)14University Hospital Carl Gustav Carus, NCT/UCC Early Clinical Trial Unit, Dresden, Germany (M. Wermke, MD)15Eli Lilly and Company, New York, NY, USA (L. Gao, PhD; D. Ferry, MD)16Eli Lilly and Company, Indianapolis, IN, USA (R. Widau, PhD; G. Mi, PhD; J. Jin, PhD)17Royal Marsden Hospital, London and Surrey, UK (I. Chau, MD)

* Correspondence to:Roy S. Herbst, MD, PhDEnsign Professor of MedicineProfessor of PharmacologyChief of Medical OncologyDirector, Thoracic Oncology Research ProgramAssociate Director for Translational ResearchYale Comprehensive Cancer CenterYale School of Medicine

333 Cedar Street, WWW221New Haven, CT 06520-8028Phone: 203-785-6879Fax: 203-737-5698

The data have not been previously published, nor are they under consideration elsewhere, but were presented in part at the 2018 ASCO Annual Meeting (Herbst RS et al. J Clin Oncol 36, 2018 [suppl; abstr 3059]).

Summary

Background: Pre-clinical and clinical evidence suggests that simultaneous blockade

of vascular endothelial growth factor receptor 2 (VEGFR-2) and programmed death 1

(PD-1) or programmed death-ligand 1 (PD-L1) enhances antigen-specific T-cell

migration, anti-tumour activity, and has favourable toxicity. In this study, we aimed to

assess the safety and preliminary efficacy of ramucirumab (IgG1 VEGFR-2

antagonist) combined with pembrolizumab (IgG4 PD-1 antagonist) in patients with

previously treated advanced gastro-oesophageal junction (G-GEJ) adenocarcinoma,

non-small-cell lung cancer (NSCLC), or urothelial carcinoma (UC).

Methods: This multi-cohort, non-randomised, open-label, Phase 1 trial enrolled adult

patients (≥18 years) with histologically confirmed G-GEJ adenocarcinoma (Cohorts A

and B), NSCLC (Cohort C), or (UC (Cohort D), whose disease progressed on one or

two lines (G-GEJ) or one to three lines (NSCLC and UC) of prior therapy that

included platinum (all tumour types) and/or fluoropyrimidine (G-GEJ), with presence

of measurable disease and Eastern Cooperative Oncology Group performance

status of 0–1. Results for the ongoing treatment-naïve G-GEJ [Cohort A2] and

NSCLC [Cohort E] cohorts will be reported separately. Pembrolizumab was

administered intravenously (IV), 200 mg, on Day 1 with IV Ramucirumab at 8 mg/kg

on Days 1 and 8 (Cohort A) or at 10 mg/kg on Day 1 (Cohorts B, C, and D), every 3

weeks until disease progression or other discontinuation criteria were met. The

primary endpoint of this study was the safety and tolerability of ramucirumab in

combination with pembrolizumab assessed by the incidence of adverse events, with

dose-limiting toxicities observed during a 21-day treatment cycle in the phase 1a

period. The safety and efficacy analysis set included all patients that received at

least 1 dose of study treatment. This trial is registered at ClinicalTrials.gov, number

NCT02443324, and is no longer enrolling patients.

Findings: Between July 30, 2015 and June 24, 2016, 92 patients were enrolled and

treated at 16 centres in five countries. Median follow-up was 32.8 months

(interquartile range [IQR] 28.1-33.6). Twenty-two (24%) patients experienced one or

more Grade ≥3 treatment-related adverse events (TRAEs), most commonly

hypertension (6 patients; 7%) and colitis (5 patients; 5%). Serious AEs (SAEs)

occurred in 53 (58%) of 92 patients, and deemed related to treatment in 22 (24%)

patients. The most commonly related SAE in the G-GEJ cohort was abdominal pain

(3 of 41 patients; 7%), in the NSCLC cohort were asthenia (2 of 27 patients; 7%),

and myocardial infarction (2 of 27 patients; 7%), and in the UC cohort was colitis (2

of 24 patients; 8%). Six (6.5%) patients discontinued because of TRAEs and one

death (pulmonary sepsis, G-GEJ cohort) was deemed related to treatment. The

objective response rates for the G-GEJ, NSCLC, and UC cohorts were 7% (95%

confidence intervals [CI] 1.5-19.9), 30%, (95% CI 13.8-50.2), and 13%, (95% CI 2.7-

32.4), respectively.

Interpretation: Ramucirumab in combination with pembrolizumab showed a

manageable safety profile with favourable anti-tumour activity in patients with

previously treated advanced G-GEJ carcinoma, NSCLC, and UC. Our results

contribute to the growing evidence that supports dual inhibition of the VEGF/VEGFR-

2 and PD-1/PD-L1 pathways. This combination could be further explored with or

without chemotherapy, especially in tumours where single-agent checkpoint

inhibitors have shown no additional benefit over chemotherapy.

Funding: Eli Lilly and Company

Research in Context

Evidence before this study

Before trial enrolment began, we searched PubMed, the abstracts of major oncology

congresses (specifically the American Society of Clinical Oncology and European

Society for Medical Oncology congresses), and ClinicalTrials.gov with the search

terms “carcinoma”, “cancer”, “immune checkpoint inhibitor”, “anti-PD-1”, “anti-PD-

L1”, “trials”, “clinical trials”, “VEGF”, and “VEGFR-2” for preclinical reports and

clinical trials published in English up to Jan 1, 2015. This search showed that

immune-checkpoint inhibitors (CPIs) targeting the PD-L1/PD-1 axis have durable

activity in various patient subsets, although many treated with CPIs had progressive

disease as their best response. There was much evidence that mechanisms of

resistance to CPIs are likely multi-factorial and can include lack of PD-L1 and/or

inhibitory effects in the tumour microenvironment. Anti-angiogenic therapies targeting

vascular endothelial growth factor (VEGF) or VEGF receptor-2 (VEGFR-2) can

increase trafficking of T cells into tumours as well as reduce immunosuppressive

cytokines and T-regulatory cells and may help overcome resistance. Pre-clinical

studies provided evidence for a synergistic antitumour effect when blocking VEGFR2

and PD-1 simultaneously in a murine colon cancer model, with dual blockade of both

PD-1 and VEGFR2 inhibiting tumour growth significantly compared to each

treatment, individually. Furthermore, in 2 studies, in patients with metastatic

colorectal cancer and melanoma, the combination of bevacizumab (anti-VEGF) and

immune checkpoint blockade shown to be well tolerated with no unexpected

toxicities.

Added value of this study

Our study combined ramucirumab (VEGFR-2 antagonist) and pembrolizumab (PD-1

antagonist) to simultaneously target both angiogenesis/tumour microenvironment

and immune checkpoint inhibition in biomarker unselected patients with previously

treated advanced NSCLC, G-GEJ or UC. The combination of ramucirumab and

pembrolizumab showed a manageable safety profile, with no indication that

ramucirumab potentiates pembrolizumab toxicity (or vice versa) and full doses of

each individual drug could be administered in combination. Efficacy endpoints in our

study compared favourably to immune-checkpoint inhibitor monotherapy in other

studies.

Implications of all the available evidence

Our results contribute to the growing evidence that supports dual inhibition of the

VEGF/VEGFR-2 and PD-1/PD-L1 pathways in patients with previously treated

advanced or metastatic cancer. Given the manageable safety profile and clinical

activity, this combination could be explored in future trials either with or without

chemotherapy, especially in tumours where single-agent CPIs have shown no

additional benefit to chemotherapy.

Introduction

Tumours can evade T cell-mediated killing through upregulation of

programmed death ligand 1 (PD-L1), which interacts with inhibitory receptor

programmed death 1 (PD-1) expressed on tumour-infiltrating T cells, leading to their

functional inactivation. Immune-checkpoint inhibitors (CPIs) targeting the PD-L1/PD-

1 axis show durable activity in a subset of patients.1-3 However, many treated with

CPIs have progressive disease as their best response and novel combination

treatments are needed. Pembrolizumab has durable anti-tumour activity in some

patients, including those with previously treated advanced non-small-cell lung cancer

(NSCLC), gastric or gastro-oesophageal junction (G-GEJ) adenocarcinoma, and

urothelial carcinoma (UC).1,2,4

Mechanisms of resistance to CPIs are likely multi-factorial and can include

lack of PD-L1 and/or inhibitory effects in the tumour microenvironment. Anti-

angiogenic therapies targeting vascular endothelial growth factor (VEGF) or VEGF

receptor-2 (VEGFR-2) can increase trafficking of T cells into tumours as well as

reduce immunosuppressive cytokines and T-regulatory cells and may help overcome

resistance.5,6 Clinical experience with anti-angiogenic agents in combination with

CPIs shows enhanced antigen-specific T-cell migration, anti-tumour activity, and

favourable toxicity.7-13 The translational work of Wallin et al. demonstrated that

bevacizumab (anti-VEGF) increases antigen-specific T-cell migration, major

histocompatibility complex 1 (MHC-1) expression, and PD-L1 expression.13 Adding

atezolizumab to bevacizumab augmented this process and leads to deep and

durable responses in metastatic renal cell carcinoma (mRCC).13 A randomised

Phase 2 trial combining bevacizumab and atezolizumab produced a superior median

progression-free survival (PFS) of 11.7 versus 8.4 months compared to sunitinib in

first-line mRCC.9 More recently, a Phase 3 study has confirmed the benefits of

adding an antiangiogenic in combination with pembrolizumab in mRCC.11

Furthermore, the boosting effects of bevacizumab on atezolizumab in combination

with chemotherapy have been seen in patients with NSCLC, demonstrating the

clinical effectiveness of combining antiangiogenics with CPIs.12 After initial treatment

with cytotoxic chemotherapy, persistent toxicities can have an impact on the ability of

many cancer patients to receive second-line therapy, and with some tumour types,

G-GEJ for example, up to 50% of patients do not receive second-line therapy.14 In

these patients there is an unmet need for possible non-chemotherapy options with

comparable activity but differing toxicity profiles.

Ramucirumab is an immunoglobulin G, subclass 1 (IgG1) monoclonal antibody that

binds to the extracellular domain of VEGFR-2 and has demonstrated anti-tumour

activity in the Phase 3 setting, as monotherapy or in combination with chemotherapy,

for multiple tumour types, including NSCLC, G-GEJ, and UC.15-18 Here, we describe

results from Study JVDF, a Phase 1a/b, open-label, multi-cohort trial of ramucirumab

with pembrolizumab for patients with advanced solid tumours. We reported results

for patients with previously treated advanced or metastatic biliary tract cancer (BTC)

separately.7 Ramucirumab plus pembrolizumab demonstrated no unexpected safety

findings in patients with BTC, but did not improve overall survival when compared

with historical controls. Results for the ongoing treatment-naïve NSCLC and G-GEJ

cohorts will be reported separately. Here, we report results for patients with

previously treated G-GEJ, NSCLC, and UC tumours.

METHODS

Study design and participants

Patients ≥18 years were enrolled at 16 centres in five countries with

histologically confirmed G-GEJ (Cohorts A and B), NSCLC (Cohort C), or UC

(Cohort D); unresectable or metastatic disease; and progression on one or two lines

(G-GEJ) or one to three lines (NSCLC and UC) of prior chemotherapy. Presence of

measurable disease was based on Response Evaluation Criteria in Solid Tumors

(RECIST) version 1.1. Prior therapy for advanced disease must have included

platinum (all tumour types) and/or fluoropyrimidine (G-GEJ). Patients with NSCLC

who have known epidermal growth factor receptor (EGFR) or anaplastic lymphoma

kinase (ALK) mutations were eligible if they had received one or more prior targeted

therapies for these mutations. An Eastern Cooperative Oncology Group performance

status (ECOG PS) of 0 or 1, measurable disease, adequate organ function

(haematologic: absolute neutrophil count ≥1.5x109/L, platelets ≥100x109/L,

haemoglobin ≥9 g/dL or ≥5.6 mmol/L; renal: creatinine ≤1.5 times the upper limit of

normal (xULN) or ≥60 mL/min; hepatic: total bilirubin ≤1.5xULN, aspartate

aminotransferase and alanine aminotransferase ≤2.5xULN or ≤5xULN for patients

with liver metastases; coagulation: International Noramlised Ratio ≤1.5 x ULN or

Prothrombin Time ≤ 5 seconds above ULN; thyroid: thyroid-stimulating hormone

within normal limits), and a newly obtained core or excisional biopsy of a tumour

lesion prior to enrolment were required. Patients were excluded if they had known

brain metastases, uncontrolled spinal cord compression, or leptomeningeal disease,

or had a serious illness or medical condition including, but not limited to,

immunodeficiency, active autoimmune disease, pneumonitis, interstitial lung disese,

hepatitis B or hepatitis C infection, human immunodeficiency virus, liver cirrhosis,

congestive heart failure. Prior therapy with agents targeting VEGF/VEGF receptor,

PD-1/PDL-1, or PD-1/PDL-2 signaling pathways, was not permitted. Patients were

not allowed to have had chemotherapy, targeted small molecule therapy, or radiation

therapy within 2 weeks and/or monoclonal antibody treatment within 4 weeks prior to

enrolment. Full eligibility criteria are provided in the appendix (pp 53-61).

The dose-limiting toxicity (DLT) observation period (Phase 1a; safety run-in)

occurred during one 21-day treatment cycle and DLTs are defined in the appendix

(Section 6.1.1.1.1). Ramucirumab had two dosing schedules (8 mg/kg Day 1 and 8,

or 10 mg/kg, Day 1), while dose of pembrolizumab was fixed at 200 mg on Day 1

every 3 weeks (Q3W).

Tumour response was assessed radiographically by the investigator

according to RECIST 1.1, with required confirmation for partial and complete

responses.

The trial adhered to the Declaration of Helsinki, the International Conference

on Harmonisation (ICH) Guidelines for Good Clinical Practice, and applicable local

regulations. The protocol was approved by ethics committees for all participating

centres, and all patients provided written informed consent before study entry.

Procedures

Patients with G-GEJ (Cohort B only), NSCLC, and UC received ramucirumab

10 mg/kg intravenously (IV) on Day 1 with pembrolizumab 200 mg IV on Day 1 Q3W.

Patients with G-GEJ in Cohort A received ramucirumab 8 mg/kg IV on Days 1 and 8

with pembrolizumab 200 mg IV on Day 1 Q3W. Treatments were continued for up to

35 cycles or until disease progression, unacceptable toxicity, or discontinuation for

any other reason.

The dosing strategy for ramucirumab is outlined in the appendix (see Section

4.2.1). A ramucirumab dose of 8 mg/kg on day 1 and 8 Q3W was based on exposure

response analyses in the REGARD (8 mg/kg Q2W) and RAINBOW (8 mg/kg on

Days 1 and 15; 28-day cycle) studies.16,17 Based on pharmacokinetic simulation, 8

mg/kg on Day 1 and 8 Q3W should produce higher ramucirumab exposure and

potentially better clinical efficacy outcomes relative to the 8-mg/kg Q2W regimen.

The ramucirumab-related safety risk was not expected to be significantly increased

as this dose is approximately 60% lower than the maximum tolerated weekly dose

identified in the Phase 1 dose-escalation (13 mg/kg weekly).

Dose modifications were permitted for non-life-threatening Grade 3 clinical

adverse events that were considered to be at least possibly related to study

treatment. Treatment was generally discontinued for Grade 4 clinical adverse events.

Further details about permitted dose modifications are in the appendix (pp 65-75). A

patient could continue with only one agent while discontinuing the other. For this

reason the duration of treatment for the drugs could be different.

Response was assessed every 6 weeks (± 7 days) for the first 24 weeks and

measured every 12 weeks (± 7 days) thereafter. Study treatment was to be

continued until progression was confirmed in a subsequent scan (after at least 4

weeks) per RECIST; once progression was confirmed treatment was discontinued.

Stable disease was confirmed if the duration was ≥39 days. Following

discontinuation, patients were followed for survival approximately every 90 days.

Safety was assessed and adverse events (AEs) were graded throughout the study

and for 30 days after treatment discontinuation. The AEs were graded using the

National Cancer Institute’s Common Terminology Criteria for Adverse Events (NCI

CTCAE), version 4.0, and judged by the investigator to be related or not related to

study treatment. Laboratory monitoring was performed within 7 days prior to cycle 1

and within 4 days prior to each subsequent cycle.

Consistent with prior methods1,19,20, PD-L1 expression was assessed using

immunohistochemistry with an investigational version of PD-L1 IHC 22C3 pharmDx

(Agilent, Carpinteria, CA, USA), for which the number of stained tumour cells

(tumour proportion score [TPS]) or tumour and immune cells (combined positive

score [CPS]) was relative to the total number of tumour cells. PD-L1 positivity was

defined by CPS≥1% in G-GEJ and UC and by TPS≥1%, 1% to 49%, and ≥50% in

NSCLC.

Objectives

The primary endpoint was the safety and tolerability of ramucirumab in

combination with pembrolizumab (dose-limiting toxicity observed during 21-day

treatment cycle, and in a Phase 1b dose-expansion period). Secondary endpoints

included PFS (by investigator review), defined as the time from date of first study

treatment until the first radiographic documentation of objective progression, or as

death due to any cause; overall survival, (OS) defined as the time from date of first

study treatment to death from any cause; objective response rate (ORR), defined as

the proportion of patients with a best overall response of complete or partial

response; disease control rate, defined as the proportion of patients with a best

overall response of complete response, partial response or stable disease; duration

of response (DOR), defined as the first date of complete or partial response until the

first date of objective progression, or death; time to response, defined as the time

from the date of first study treatment until the first evidence of a confirmed complete

or partial response; and pharmacokinetics of ramucirumab. Exploratory endpoints

include immunogenicity of ramucirumab when co-administered with pembrolizumab;

and association between biomarkers with clinical outcomes.

Statistical analysis

Safety and efficacy analysis sets were based on all enrolled patients.

Patients enrolled in the trial were those who had been assigned to a treatment and

had received at least 1 dose of study treatment.

Planned enrolment for each tumour expansion cohort was 25 to 30 patients.

Fifteen patients each were planned for both G-GEJ Cohorts (A and B). The sample

size was selected to allow adequate assessment of safety and preliminary efficacy at

the recommended doses for ramucirumab and pembrolizumab. For the G-GEJ,

NSCLC and UC cohorts, the null hypothesis was based on the assumption that the

ORR is no greater than 10% to 15% and the target treatment effect of the

combination therapy on ORR is between 20% and 30%. Based on these response

rates, a sample size of 25 to 30 patients per cohort (n=30 for combined gastric

Cohorts A and B) provided approximately 60% to 90% power with a one-sided α

level of 0.20. The 95% confidence intervals (CI) for the ORR were calculated based

on the Clopper-Pearson method. Patients who did not have any postbaseline tumour

response assessments for any reason were considered non-evaluable and were

included in the denominator when calculating the response rate. Time-to-event

variables were estimated using Kaplan-Meier methodology. Safety was assessed in

all patients who received one or more doses of study medication. The time to

resolution of an AE is calculated from the start date of the first episode of the AE,

until the resolved/recovered date for the same episode. For the Phase 1a part of the

study, the available safety and pharmacokinetic data were reviewed on a cohort-by-

cohort basis. For the Phase 1b part of the study, interim analyses occurred at a

cohort level when the patients had completed approximately 24 weeks of study

treatment or discontinued for any reason. SAS® 9.4 and R version 3.5.0 were used

for statistical analyses.

The protocol was amended (February 11, 2016) to include 3 new phase 1b

expansion cohorts, first-line G/GEJ (cohort A2), first-line NSCLC (cohort E), and

second or third line biliary tract cancer (cohort A1). The rationale for this amendment

is described in the protocol (appendix pp 146-182); (results for the ongoing cohorts

A2 and E will be reported separately).

This trial is registered with ClinicalTrials.gov, number NCT02443324.

Role of the funding source

The study funder (Eli Lilly) had roles in study design, data collection, and

analysis. The funder interpreted data in collaboration with all authors and supported

development of the report by providing writing and editorial assistance. The

corresponding author had full access to all study data and all authors had final

responsibility for the decision to submit for publication.

RESULTS

Between July 30, 2015 and June 24, 2016, 138 patients were screened, of

whom 46 were excluded (figure 1); 92 eligible patients were assigned to receive

study treatment in G-GEJ (A, n=24; B, n=17), NSCLC (n=27), or UC (n=24) cohorts.

Baseline demographics and disease characteristics are summarised in table 1 (prior

therapy shown in appendix pp 3-4).

Data cut-off for the current analysis was August 31, 2018, at which time all 92

(100%) patients had discontinued treatment. The median follow-up duration was

32.8 months (interquartile range [IQR], 28.1-33.6).

All patients received one or more doses of ramucirumab and pembrolizumab

(appendix p 5). The median therapy duration for patients with G-GEJ was 2.8

months (IQR, 1.4-5.5) with ramucirumab and 3 months (IQR, 1.4-7.6) with

pembrolizumab. In NSCLC, the median therapy duration was 7 months (IQR, 3.0-

16.8) with ramucirumab and 8.3 months (IQR, 3.3-23.7) with pembrolizumab. In UC,

the median therapy duration was 2.1 months (IQR, 1.4-4.6) with ramucirumab and

2.4 months (IQR, 1.5-4.6) with pembrolizumab. The median relative dose intensities

were ≥95% for ramucirumab and pembrolizumab in all patients. Ramucirumab dose

was reduced in 2 patients in the G-GEJ cohort and 1 patient in the UC cohort. No

dose reductions in pembrolizumab were required. Per the data cut-off, 9 (22%)

patients with G-GEJ, 11 (41%) with NSCLC, and 6 (25%) with UC received post-

study therapy (appendix p 6). Of note, one patient with G-GEJ (Cohort B) and 7

patients with NSCLC discontinued study after completing the maximum duration of

study treatment (35 cycles).

The DLTs were assessed during the first 21 days of study treatment in the

Phase 1a portion of the study. During the first cycle of treatment, one patient with

GEJ adenocarcinoma (ramucirumab 8 mg/kg Days 1 and 8 Q3W, in combination of

pembrolizumab 200 mg IV Day 1 Q3W) experienced Grade 3 abdominal pain, colitis,

hepatitis, interstitial lung disease, and jaundice, and Grade 4 cholestasis. This

patient died on treatment (Day 40) and the death was deemed related to progressive

disease. No additional DLTs were observed and the decision was made to maintain

the full planned doses of ramucirumab and pembrolizumab in Phase 1b. Neither

drug was escalated/de-escalated since no DLTs were observed at initial doses.

In the combined Phase 1a/b cohorts, TRAEs occurred in 75 (82%) of 92

patients. The most frequently reported TRAEs (any grade, occurring in >10% of

patients) were fatigue (33 of 92 patients [36%]), hypertension (16 [17%]),

hypothyroidism (13 [14%]), nausea (11 [12%]), and decreased appetite (10 [11%]).

These occurred predominantly at Grade 1 or 2 severity (table 2). The TRAEs of

Grades ≥3 occurred in 22 (24%) of 92 patients; most commonly hypertension (6

[7%]) and colitis (5 [5%]). One patient (cohort A) died due to pulmonary sepsis,

deemed related to treatment (the trial was not designed to distinguish attribution to

any individual study drug). Including events related by the investigator to disease

progression, deaths on treatment or within 30 days of discontinuation (any causality)

were reported for 14 (34%) of 41 G-GEJ patients (10 study disease; 1

haematemesis; 1 pulmonary sepsis; 1 sudden death; 1 bowel obstruction), 3 (11%)

of 27 NSCLC patients (3 study disease), and 8 (33%) of 24 UC patients (7 study

disease; 1 thromboembolic event). Serious AEs (SAEs) were reported for 53 (58%)

of 92 patients; these were deemed related to treatment in 22 (24%) patients. The

most commonly related SAE in the G-GEJ cohort was abdominal pain (3 of 41

patients; 7%), in the NSCLC cohort were asthenia (2 of 27 patients; 7%), and

myocardial infarction (2 of 27 patients; 7%), and in the UC cohort was colitis (2 of 24

patients; 8%) (appendix p 7). Six (6.5%) patients discontinued treatment due to

TRAEs (Grade 1 diarrhoea; Grade 2 fatigue; and Grade 3 asthenia, myocardial

infarction, pulmonary embolism, and colitis).There were no obvious differences in

safety between the 2 different schedules of ramucirumab used in the G-GEJ cohorts

(Cohorts A and B). TRAEs were reported in 21 (88%) of 24 patients in Cohort A, and

13 (76%) of 17 patients in Cohort B; TRAEs of Grades ≥3 occurred in 7 (29%)

patients and 4 (24%) patients, respectively.

The AEs of special interest, based on the known safety profile of

ramucirumab and pembrolizumab, are shown inthe appendix (pp 8-11) . Based on

the recently published Trial Reporting in Immuno-oncology (TRIO)

recommendations,21 additional information on immune-related AE dose delay, dose

discontinuation, timing and duration of toxicity onset, and hospitalisations are shown

in the appendix (pp 12-13) . One fatal event of embolism occurred in a patient with

UC (4%), 1 fatal event of gastrointestinal haemorrhage causing haematemesis

occurred in a patient with GEJ (Cohort B), and 1 fatal event of intestinal obstruction

occurred in a patient with GEJ (Cohort B); all 3 were deemed unrelated to study

treatment.

In combined G-GEJ cohorts, at the time of data cut-off, 38 (93%) of 41

patients had disease progression of which 33 (80%) patients had died. Confirmed

partial responses occurred in 3 (7%) of 41 patients, 18 (44%) patients had stable

disease, and 13 (32%) had progressive disease as best response to treatment

(figure 2A and table 3). The median time to response (TTR) was 1.4 months (95%

CI, 1.4-4.1) and median DOR was 6.7 months (95% CI, 4.4-17.5; appendix p 18).

Disease control occurred in 21 (51%) patients, with a median duration of stable

disease (DOSD) of 5.0 months (95% CI 4.0-8.5; table 3). The median PFS was 2.5

months (95% CI, 1.5-4.2) and the estimated PFS rate at 3 and 6 months was 44%

and 26%, respectively (table 3 and figure 2B). The median OS was 5.9 months (95%

CI, 4.4-10.6) and the estimated OS rate at 6 and 12 months was 49% and 31%,

respectively (table 3 and figure 2C). Visual inspection of the survival curves and the

response rates did not suggest any differences in efficacy between the two dose

schedules used in G-GEJ patients (safety profiles were also similar). Patients with

PD-L1–positive tumours (n=22) had improved efficacy results compared to patients

with PD-L1–negative disease (n=17), including median OS of 12.6 months (4.7-20.3)

and 5.2 months (1.3-8.6), respectively (appendix pp 14, 21). One patient

discontinued study treatment after completing the maximum number of 35 cycles of

treatment per the protocol (figure 2A).

In the NSCLC cohort, at the time of data cut-off, 17 (63%) of 27 patients had

disease progression of which 15 (56%) patients had died. Confirmed objective

responses occurred in 8 (30%) of 27 patients. One (4%) patient had a complete

response, 7 (26%) patients had a partial response, 15 (56%) had stable disease, and

3 (11%) had progressive disease as their best response to treatment (figure 3 and

table 3). The median TTR was 2.1 months (95% CI, 1.3-3.0). Responses were

durable and occurred regardless of PD-L1 status, with a median DOR not reached

(NR) (95% CI 11.1 months -NR; table 3 and appendix pp 14, 19). Disease control

occurred in 23 (85%) patients, with a median DOSD of 6.9 months (95% CI, 2.8-9.7;

table 3). The median PFS was 9.7 months (95% CI, 4.6-27.6) and the estimated PFS

rate at 6 and 12 months was 65% and 43%, respectively (table 3; figure 3B). The

median OS was 26.2 (95% CI, 11.8-NR) months and the estimated OS rate at 12

and 18 months was 68% and 59%, respectively (table 3; figure 3C). Efficacy results

by PD-L1 expression are reported in the appendix (pp 14, 22). One patient with

NSCLC was positive for EGFR at baseline and prior treatment included erlotinib.

This patient was PD-L1 positive, had a best overall response to study treatment of

stable disease, and discontinued at cycle 9 due to progressive disease. Seven (26%)

of 27 patients discontinued study treatment after completing the maximum number of

35 cycles of treatment per the protocol (appendix p 19).

In the UC cohort, at the time of data cut-off, 21 (88%) of 24 patients had

disease progression of which 18 (75%) patients had died. Confirmed partial

responses occurred in 3 (13%) of 24 patients, 9 (38%) patients had stable disease,

and 11 (46%) had progressive disease as their best response to treatment (figure 4A

and table 3); of note, confirmed responses only occurred in PD-L1 positive patients.

The median DOR was 8.3 months (4.6-16.8) and median TTR was 2.8 months (1.3-

5.5; appendix p 20). Disease control occurred in 12 (50%) patients with a median

DOSD of 2.8 months (95% CI 1.9-13.1; table 3). The median PFS was 1.9 months

(95% CI, 1.2-2.8) and the estimated PFS rate at 3 and 6 months was 28% and 23%,

respectively (table 3; figure 4B). The median OS was 6.4 months (95% CI, 2.5-18.7)

and the estimated OS rate at 6 and 12 months was 52% and 38%, respectively

(table 3; figure 4C). Efficacy results by PD-L1 expression are reported in the

appendix (pp 14, 23).

Ramucirumab exposure levels appeared to be similar in patients with

different cancer indications when ramucirumab was administered following the same

schedule. Dose-related ramucirumab exposure increases were observed between 8

mg/kg Days 1 and 8 Q3W and 10 mg/kg Q3W regimens. As expected, 8 mg/kg Days

1 and 8 Q3W regimen produced higher trough concentrations, but comparable peak

concentration, relative to 10 mg/kg Q3W regimen (appendix pp 15-17). These data

are consistent with those from previous studies in which ramucirumab was

administered to patients with various types of cancer using same regimens.

DISCUSSION

We report that in 92 patients with advanced or metastatic G-GEJ (second- to

third-line), NSCLC (second- to fourth-line), and UC (second- to fourth-line),

ramucirumab in combination with pembrolizumab revealed no unexpected safety

findings. The most common toxic effects were of Grade 1 or 2 severity and

manageable with supportive care or with dose reduction/delay, with a low

discontinuation rate due to TRAEs. Grade 3 TRAEs were experienced by 22 (24%)

patients, the most common being hypertension (7%). Four (4%) patients

experienced Grade 4 TRAEs and 1 patient died due to pulmonary sepsis, deemed

related to treatment. The majority of SAEs were Grade 3 or lower and the proportion

of patients experiencing an SAE were similar or lower to those previously reported

for monotherapy ramucirumab or pembrolizumab.1,16 The majority of these events

were recovered, resolved, or were resolving; six (6.5%) patients discontinued

treatment due to treatment-related AE. The rates of certain AEs identified as

potential class effects of antiangiogenic therapies, such as hypertension and

bleeding (mainly Grade 1 or 2 epistaxis), were in the range observed in a meta-

analysis of 4996 patients of randomised ramucirumab trials.22 Likewise, the rates of

all-grade arterial (3%) and venous (8%) thromboembolic events were similar to the

frequencies reported in the late-phase meta-analysis of 1.4% and 3.9%,

respectively.22 Rates of AEs identified as potential class effects of CPIs, such as

colitis, hypothyroidism, and pneumonitis, were relatively uncommon and mainly

occurred at Grade 1 or 2 severity. Events of Grade 3 colitis (5%) and pneumonitis

(1%) were in the range reported by meta-analyses for CPIs.23,24 Acknowledging the

limitations of the sample size, we have no indication that ramucirumab potentiates

pembrolizumab toxicity (or vice-versa) and full doses of each individual drug could

be administered in combination (table 2,appendix pp 8-11).

Different dose schedules were used in the G-GEJ cohorts (Cohorts A and B),

because when this trial started, significant questions remained about the optimal

dose of ramucirumab. Subsequent dose-response studies have demonstrated that

higher doses of ramucirumab have not proven to provide significantly improved

clinical benefits compared to the approved regimen.25 There was no observed safety

or efficacy difference between the 2 different schedules of ramucirumab used in G-

GEJ patients.

Expression of PD-L1 on tumour and immune cells is associated with

increased clinical benefit from PD-1/PD-L1 inhibitors in multiple tumour types.3 We

did not restrict enrolment based on PD-L1 status and only half (49%) of the tested

biopsies were positive for PD-L1. Similar to prior reports, efficacy was generally

better in PD-L1–positive patients and this was most evident in patients with G-GEJ

(second- to third- line). However, durable anti-tumour activity was also observed in

patients who were negative for PD-L1, most evident in patients with NSCLC. Future

randomised trial designs should consider stratification by PD-L1 status.

In patients with G-GEJ in the ATTRACTION-2 trial, the clinical activity of

nivolumab (anti–PD-1) monotherapy versus placebo was observed in a third-line and

subsequent line trial of patients, with a PFS of 1.6 months and a 12-month OS rate

of 26.2%.26 Likewise, activity of pembrolizumab monotherapy was demonstrated in a

single-arm cohort of largely third-line cases (KEYNOTE-059), with a PFS of 2.0

months and a 12-month OS rate of 23.4% (PD-L1–positive patients had a median

OS of 5.8 months).27 Recent data from the Phase 3 KEYNOTE-061 trial showed that

pembrolizumab was not better than paclitaxel alone in second-line G-GEJ patients

with PD-L1–positive tumours, with the PFS reported for pembrolizumab as 1.5

months and 12-month OS rate of approximately 40%.19 Response to pembrolizumab

in G-GEJ decreases with later lines of therapy, perhaps because patients may be

less likely to manifest an immune response after multiple lines of therapy.27 The

combination of ramucirumab and pembrolizumab may help overcome this decrease

in activity of a CPI in later lines of therapy, and it may be even more effective in

earlier lines of therapy. Here, we report a PFS of 4.6 months, OS of 12.6 months,

and 12-month OS rate of 55% in the PD-L1–positive cases, which were 59% third-

line. Seven (17%) of G-GEJ patients were not-evaluable for response. This may be

due to chance, with a larger proportion of patients having had clinical disease

progression prior to radiological evaluation, thus not being evaluable, although there

is no reason this proportion should be larger than the NSCLC and UC cohorts.

The CPIs have shown significant benefit in patients with NSCLC.28 The

activity of pembrolizumab monotherapy was first demonstrated in a large single-arm

Phase 1 study of PD-L1 all comers in KEYNOTE-001, with a PFS of 3.0 months,

ORR of 18%, and OS of 9.3 months in patients previously treated for NSCLC.29

Broadly, in the second-line setting, the activity of pembrolizumab monotherapy was

superior to docetaxel in patients with PD-L1–positive tumours (KEYNOTE-010), with

an ORR of 18%, PFS of 3.9 months, and median OS of 10.4 in patients receiving the

2-mg/kg dose.2 The activity of pembrolizumab monotherapy correlates with PD-L1

expression, with the most favourable clinical outcomes in patients with PD-L1

expression on ≥50% of tumour cells.2,20 Here, we report an ORR of 30%, PFS of 9.7

months, and median survival of 26.2 months in a PD-L1–all-comer population, which

consisted of 44% third-line cases. The interpretive challenges of cross-trial

comparisons notwithstanding, efficacy end points in our study compare favourably to

CPI in other studies, underlining the potential activity of this regimen in biomarker-

unselected patients.2,28-30

Ramucirumab plus docetaxel improves PFS (median 4.1 months vs. 2.8

months) and ORR (24.5% vs. 14%) compared to placebo plus docetaxel in patients

with platinum-refractory advanced UC.18 Inconsistencies have been seen with CPIs

in platinum-refractory UC, although a survival advantage has been shown for

pembrolizumab when compared to docetaxel.1 Here, we report an ORR of 13%, PFS

of 1.9 months, and OS of 6.4 months in PD-L1 all comers, who were 50% third-line

patients, all with visceral metastases at baseline, and 13 (54%) were high-risk by

Bellmunt scoring. Patients who were PD-L1–positive had better outcomes, including

a tail in the OS curve. No responses were seen in PD-L1–negative patients.

Our results contribute to the growing evidence that supports dual inhibition of

the VEGF/VEGFR-2 and PD-1/PD-L1 pathways in patients with previously treated

advanced or metastatic cancer.7-13 Observed efficacy results in the NSCLC cohort

were particularly striking when compared with the other tumour types in this study,

which could be driven by the robust efficacy results in both PD-L1-negative and PD-

L1-positive NSCLC tumours. Efficacy results in G/GEJ and UC were mainly seen in

PD-L1-positive tumours.

Limitations of our study include the relatively small size of the expansion

cohorts. We did not include p-values as the study is not designed to have sufficient

power for subgroup analyses (by PD-L1 status); the even smaller sample size from

those subgroups makes p-value non-informative. A greater understanding of tumour

biology and its interaction with the immune system is needed in patients treated with

antiangiogenic agents in combination with immune checkpoint inhibitors. Our study

was limited in providing greater understanding of this complex interaction because

the protocol did not allow for collection of post- treatment biopsies. In addition, single

treatment arm design of the study limited our ability to directly compare findings with

other available therapies in each tumour type studied.

Given the manageable safety profile and clinical activity, this combination

could be explored in future trials either with or without chemotherapy, especially in

tumours for which single-agent CPIs have failed to show benefit over chemotherapy.

Author contributions

RSH, CF, DPP, and IC contributed to study design, data collection, data

interpretation, and drafting, review and approval of the submitted report. HTA, RSD,

EC, LPA, PAC, JB, NP, MGK, JML, NI, AS, MW, and JC contributed to data

collection, data interpretation, and drafting, review and approval of the submitted

report. GM, DF, and LG contributed to study design, data analysis, data

interpretation, and drafting, review and approval of the submitted report. RCW and

JJ contributed to data analysis, data interpretation, and drafting, review and approval

of the submitted report.

Acknowledgements

This study is sponsored by Eli Lilly and Company, in collaboration with

Merck and Co., Inc. We thank the patients, their families, and the study personnel

across all sites for participating in this study. Writing assistance was provided by

John Bilbruck of ProScribe, part of the Envision Pharma Group, who was funded by

Eli Lilly and Company. Editorial assistance was provided by Atasi Laha and Purnima

Chandra, employees of Eli Lilly and Company.

Declaration of interest

R.S.H has received honoraria from for following: Consulting: Abbvie

Pharmaceuticals, ARMO Biosciences, AstraZeneca, Biodesix, Bristol-Myers Squibb,

Eli Lilly and Company, EMD Serrano, Genentech/Roche, Genmab, Halozyme, Heat

Biologics, Infinity Pharmaceuticals, Loxo Oncology, Merck and Company, Nektar,

Neon Therapeutics, NextCure, Novartis, Pfizer, Sanofi, Seattle Genetics, Shire PLC,

Spectrum Pharmaceuticals, Symphogen, Tesaro, Tocagen. Research support:

AstraZeneca, Eli Lilly and Company, Merck and Company. He is a member of the

board of directors (non-executive/ independent) for Junshi Pharmaceuticals.

HTA reports no conflicts outside the submitted work.

RSD reports research grants from Eli Lilly and Company.

EC reports research grants from AbbVie, Amcure, Amgen, AstraZeneca, BMS,

Boehringer-Ingelheim, CytomX, Eli Lilly and Company, H3, Incyte, Kura, LOXO,

Macrogenics, Menarini, Merc, Merck Serono, Merus, Millenium, Nanobiotix, Janssen,

Nektar, Novartis, Pfizer, PharmaMar, Principia Bayer, PsiOxus, PUMA, Rigontec,

Roche/Genentech, Sanofi, Taiho, Tearo; consulting fees from AbbVie, Amcure,

AstraZeneca, Boehringer Ingelheim, Celgene, Cerulean Pharma, EUSA, GLG,

Guidepoint Global, Janssen-Cilag, Nanobiotix, Janssen, Novartis, Pfizer, Pierre

Fabre, PsiOxus Therapeutics, Roche/Genentech, Seattle Genetics, Servier;

speakers’ bureau work for Novartis; employment/ownership/leadership of START;

employment and honoraria from HM Hospitals Group; President and Founder of

NPO Foundation Intheos (Investigational Therapeutics in Oncological Sciences).

LPA reports advisory fees from Roche, Eli Lilly and Company, MSD, Bristol-Myers

Squibb, AstraZeneca, Boehringer Ingelheim, Pfizer, Amgen, Takeda, Pharma Mar,

Novartis, Celgene, outside the submitted work.

PC reports personal fees and non-financial support from AstraZeneca, non-financial

support from Roche; grants, personal fees and non-financial support from MSD, non-

financial support from Plexxikon, grants and non-financial support from Novartis.

JB reports grants from AbbVie, Acerta Pharma, ADC, Agios, Amgen, Apexigen, Arch

Oncology, ARMO, Array, Arrys, AstraZeneca, Bayer, BI, Blueprint, BMS, Boston

Biomedical, Calithera, Celgene, Celldex, Cytomx, Daiichi Sankyo, Effector, Eisai,

EMD Serono, Evelo, Five Prime, FORMA, Forty Seven, Genentech/Roche, Gilead,

GSK, Harpoon, Imclone, Incyte, Innate, Ipsen, Jacobio, Koltan, LEAP, Lilly,

Macrogenics, Marshall Edwards, MedImmune, Merck, Merrimack, Mersana, Merus,

Millennium, Nektar, Novartis, Novocare, Oncogenex, OncoMed, Onyx, Pfizer, Pieris,

Prelude Oncology, Regenix, Sanofi, Sierra, SynDevRex, Taiho, Takeda, Tarveda,

TG Therapeutics, Tracon, Tyrogenex, Unum Therapeutics, Vyriad; consulting

services for Amgen, Apexigen, Arch Oncology, ARMO, Array, AstraZeneca, Bayer,

Beigene, BMS, Celgene, Continuum Clinical, Cyteir, Daiichi Sankyo, Five Prime,

FORMA, Genentech / Roche, Gilead, GSK, Incyte, Innate, Ipsen, Janssen, LEAP,

Lilly, Macrogenics, MedImmune, Merck, Merrimack, Modern Therapeutics, Molecular

Partners, Novartis, Oncogenex, OncoMed, Phoenix Bio, Prelude Therapeutics,

Sanofi, Seattle Genetics, Taiho, Tanabe Research Laboratories, TD2 (Translational

Drug Development), TG Therapeutics, Tizona, Tolero, Torque; personal fees from

ARMO, BI, BMS, Celgene, FORMA, Genentech / Roche, Gilead, Ipsen, Lilly,

MedImmune, Merck, Novartis, Oncogenex, OncoMed, Taiho.

NP reports no conflicts outside the submitted work.

MK reports personal fees from Roche, Janssen, Achilles Therapeutics, Octimet,

AstraZeneca, BerGenBio, non-financial support from MSD, grants and non-financial

support from BerGenBio, outside the submitted work.

JML reports personal fees from Astellas, Bristol-Myers Squibb, MSD, Novartis,

Pierre Fabre, Pfizer Roche; other funding from Astellas, Bristol-Myers Squibb, Ipsen,

MSD, Novartis, Pierre Fabre, Pfizer, Roche during the conduct of the study.

NI reports no conflicts outside the submitted work.

AS reports no conflicts outside the submitted work

MW reports personal fees from Roche, non-financial support from Bristol-Myers

Squibb, personal fees from Merck and Novartis outside the submitted work.

JC reports speaker bureau participation for Celgene, Spectrum and Amgen; acting

as consultant and on advisory boards for Spectrum.

LG is an employee of Eli Lilly and Company.

RW reports employment and ownership interest from Eli Lilly and Company, during

the conduct of the study.

GM is an employee of Eli Lilly and Company.

JJ is an employee of Eli Lilly and Company.

DF is an employee of Eli Lilly and Company.

CF reports consulting role for Agios, Bain Capital, Bayer, Celgene, Dicerna, Five

Prime Therapeutics, Gilead Sciences, Eli Lilly and Company, Entrinsic Health,

Genentech, KEW, Merck, Merrimack Pharmaceuticals, Pfizer, Sanofi, Taiho, and

Unum Therapeutics. He also serves as a Director for CytomX Therapeutics and

owns unexercised stock options for CytomX and Entrinsic Health.

DP reports consultant fees from Ada Cap, Amgen, Astellas, AstraZeneca, Bayer,

Boehringer Ingelheim, Bristol-Meyers Squibb, Clovis, Eli Lilly and Company, Exelixis,

Incyte, Janseen, Pfizer, Pharmacyclics, Roche, Seattle Genetics, Urogen; grant

support from Ada Cap, Amgen, Astellas, AstraZeneca, Bayer , Bristol-Meyers

Squibb, Clovis, Eli Lilly and Company, Endocyte, Genentech, Innocrin, MedImmune,

Merck, Novartis, Pfizer, Progenics, Roche, Sanofi Aventis, Seattle Genetics;

ownership interest/investment in Bellicum/Tyme.

IC reports personal fees for advisory boards from AstraZeneca, Bayer, Bristol-

Meyers Squibb, Eli Lilly and Company, Merck-Serono, MSD, Oncologie International,

Roche; research grants from Eli Lilly and Company, Janssen-Cilag, Merck-Serono,

Sanofi Oncology; honoraria from Five Prime Therapeutics.

Data sharing statement

Eli Lilly provides access, after anonymisation, to all individual participant data

collected during the trial, except for pharmacokinetic and genetic data. Data can be

requested 6 months after the indication studied has been approved in the USA and

EU or after primary publication acceptance, whichever is later. No expiration date for

data requests is set once the data are made available. Access is provided after a

proposal has been approved by an independent review committee identified for this

purpose, and after receipt of a signed data-sharing agreement. Data and documents,

including the study protocol, statistical analysis plan, clinical study report, and blank

or annotated case report forms, will be provided in a secure data-sharing

environment for up to 2 years per proposal. Further details about submitting a data

request are available online (https://www.clinicalstudydatarequest.com).

References

1. Bellmunt J, de Wit R, Vaughn DJ, et al. Pembrolizumab as Second-Line Therapy for

Advanced Urothelial Carcinoma. N Engl J Med 2017; 376: 1015-26.

2. Herbst RS, Baas P, Kim D-W, et al. Pembrolizumab versus docetaxel for previously

treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a

randomised controlled trial. Lancet 2016; 387: 1540-50.

3. Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and immune correlates of

anti-PD-1 antibody in cancer. N Engl J Med 2012; 366: 2443-54.

4. Muro K, Chung HC, Shankaran V, et al. Pembrolizumab for patients with PD-L1-

positive advanced gastric cancer (KEYNOTE-012): a multicentre, open-label, phase 1b

trial. Lancet Oncol 2016; 17: 717-26.

5. Finke JH, Rini B, Ireland J, et al. Sunitinib reverses type-1 immune suppression and

decreases T-regulatory cells in renal cell carcinoma patients. Clin Cancer Res 2008;

14: 6674-82.

6. Terme M, Pernot S, Marcheteau E, et al. VEGFA-VEGFR pathway blockade inhibits

tumor-induced regulatory T-cell proliferation in colorectal cancer. Cancer Res 2013;

73: 539-49.

7. Arkenau HT, Martin-Liberal J, Calvo E, et al. Ramucirumab Plus Pembrolizumab in

Patients with Previously Treated Advanced or Metastatic Biliary Tract Cancer:

Nonrandomized, Open-Label, Phase I Trial (JVDF). Oncologist 2018.

8. Golan T, Lin C-C, Fu S, et al. PD-010A multi-cohort phase 1 study of ramucirumab

plus durvalumab: Preliminary safety and clinical activity in patients with locally

advanced and unresectable or metastatic gastric or gastroesophageal junction

adenocarcinoma. Ann Oncol 2017; 28: mdx263.009-mdx263.009.

9. McDermott DF, Huseni MA, Atkins MB, et al. Clinical activity and molecular correlates

of response to atezolizumab alone or in combination with bevacizumab versus

sunitinib in renal cell carcinoma. Nat Med 2018; 24: 749-57.

10. Motzer RJ, Penkov K, Haanen J, et al. Avelumab plus Axitinib versus Sunitinib for

Advanced Renal-Cell Carcinoma. N Engl J Med 2019.

11. Rini BI, Plimack ER, Stus V, et al. Pembrolizumab plus Axitinib versus Sunitinib for

Advanced Renal-Cell Carcinoma. N Engl J Med 2019.

12. Socinski MA, Jotte RM, Cappuzzo F, et al. Atezolizumab for First-Line Treatment of

Metastatic Nonsquamous NSCLC. N Engl J Med 2018; 378: 2288-301.

13. Wallin JJ, Bendell JC, Funke R, et al. Atezolizumab in combination with bevacizumab

enhances antigen-specific T-cell migration in metastatic renal cell carcinoma. Nat

Commun 2016; 7: 12624.

14. Kanagavel D, Fedyanin M, Tryakin A, et al. Second-line treatment of metastatic gastric

cancer: Current options and future directions. World J Gastroenterol 2015; 21: 11621-

35.

15. Garon EB, Ciuleanu TE, Arrieta O, et al. Ramucirumab plus docetaxel versus placebo

plus docetaxel for second-line treatment of stage IV non-small-cell lung cancer after

disease progression on platinum-based therapy (REVEL): a multicentre, double-blind,

randomised phase 3 trial. Lancet 2014; 384: 665-73.

16. Fuchs CS, Tomasek J, Yong CJ, et al. Ramucirumab monotherapy for previously

treated advanced gastric or gastro-oesophageal junction adenocarcinoma (REGARD):

an international, randomised, multicentre, placebo-controlled, phase 3 trial. Lancet

2014; 383: 31-9.

17. Wilke H, Muro K, Van Cutsem E, et al. Ramucirumab plus paclitaxel versus placebo

plus paclitaxel in patients with previously treated advanced gastric or gastro-

oesophageal junction adenocarcinoma (RAINBOW): a double-blind, randomised

phase 3 trial. Lancet Oncol 2014; 15: 1224-35.

18. Petrylak DP, de Wit R, Chi KN, et al. Ramucirumab plus docetaxel versus placebo plus

docetaxel in patients with locally advanced or metastatic urothelial carcinoma after

platinum-based therapy (RANGE): a randomised, double-blind, phase 3 trial. Lancet

2017.

19. Shitara K, Ozguroglu M, Bang YJ, et al. Pembrolizumab versus paclitaxel for

previously treated, advanced gastric or gastro-oesophageal junction cancer

(KEYNOTE-061): a randomised, open-label, controlled, phase 3 trial. Lancet 2018.

20. Reck M, Rodriguez-Abreu D, Robinson AG, et al. Pembrolizumab versus

Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer. N Engl J Med 2016;

375: 1823-33.

21. Tsimberidou AM, Levit LA, Schilsky RL, et al. Trial Reporting in Immuno-Oncology

(TRIO): An American Society of Clinical Oncology-Society for Immunotherapy of

Cancer Statement. J Clin Oncol 2019; 37: 72-80.

22. Arnold D, Fuchs CS, Tabernero J, et al. Meta-analysis of individual patient safety data

from six randomized, placebo-controlled trials with the antiangiogenic VEGFR2-binding

monoclonal antibody ramucirumab. Ann Oncol 2017; 28: 2932-42.

23. Abdel-Rahman O, Fouad M. Risk of pneumonitis in cancer patients treated with

immune checkpoint inhibitors: a meta-analysis. Ther Adv Respir Dis 2016; 10: 183-93.

24. Wang DY, Ye F, Zhao S, et al. Incidence of immune checkpoint inhibitor-related colitis

in solid tumor patients: A systematic review and meta-analysis. OncoImmunology

2017; 6: e1344805.

25. Ajani JA UA, Sarosiek T, Schenker M, Morgan C, Pikiel J, Joseph M, Salek T,

Tournigand C, Ferry DR, Zhang Y, Long A, Kuo W-L, Gao L, Russo F, Mansoor W.

Ramucirumab treatment in patients with gastric cancer/gastroesophageal junction

adenocarcinoma: Secondary analysis of efficacy and safety results of 4 dosing

regimens in the phase II trial I4T-MC-JVDB. Journal of Clinical Oncology 2018; 36:

abstr 117.

26. Kang YK, Boku N, Satoh T, et al. Nivolumab in patients with advanced gastric or

gastro-oesophageal junction cancer refractory to, or intolerant of, at least two previous

chemotherapy regimens (ONO-4538-12, ATTRACTION-2): a randomised, double-

blind, placebo-controlled, phase 3 trial. Lancet 2017; 390: 2461-71.

27. Fuchs CS, Doi T, Jang RW, et al. Safety and Efficacy of Pembrolizumab Monotherapy

in Patients With Previously Treated Advanced Gastric and Gastroesophageal Junction

Cancer: Phase 2 Clinical KEYNOTE-059 Trial. JAMA Oncol 2018; 4: e180013.

28. Herbst RS, Morgensztern D, Boshoff C. The biology and management of non-small

cell lung cancer. Nature 2018; 553: 446-54.

29. Garon EB, Rizvi NA, Hui R, et al. Pembrolizumab for the Treatment of Non–Small-Cell

Lung Cancer. N Engl J Med 2015; 372: 2018-28.

30. Borghaei H, Paz-Ares L, Horn L, et al. Nivolumab versus Docetaxel in Advanced

Nonsquamous Non–Small-Cell Lung Cancer. N Engl J Med 2015; 373: 1627-39.

Table 1: Baseline demographics and characteristics

  G-GEJn=41

NSCLCn=27

UCn=24

Median age, years (range) 58 (24-78) 65 (34-82) 63 (45-87)

Male Sex 31 (76) 21 (78) 14 (58)

Ethnic origin

White 35 (85) 26 (96) 22 (92)

Asian 4 (10) 0 0

Black or African American 1 (2) 1 (4) 0

Missing 1 (2) 0 2 (8)

ECOG performance status

0 13 (32) 7 (26) 12 (50)

1 27 (66) 20 (74) 11 (46)

2* 1 (2) 0 1 (4)

PD-L1 status

Positive 22 (54)

TPS 1% to 49%: 4

(15)

TPS≥50%: 7 (26)

12 (50)

Negative 17 (41) 11 (41) 11 (46)

Not reported 2 (5) 5 (19) 1 (4)

Histopathological grade

Low grade 1 (2) 4 (15) 2 (8)

Intermediate grade 8 (20) 1 (4) 2 (8)

High grade 24 (59) 8 (30) 13 (54)

Unable to determine 7 (17) 14 (52) 7 (29)

Not reported 1 (2) 0 0

Smoking status

Former or current 21 (51) 26 (96) 17 (71)

Never 19 (46) 1 (4) 7 (29)

Unknown 1 (2) 0 0

Previous systemic therapies

Adjuvant therapy 10 (24) 4 (15) 7 (29)

Neo-adjuvant therapy 10 (24) 2 (7) 4 (17)

Prior systemic therapy for advanced disease#

1 prior line 17 (41) 13 (48) 9 (38)

2 prior lines 24 (59) 12 (44) 12 (50)

3 or more prior lines 0 2 (7) 3 (13)

G-GEJ

GEJ as primary site 25 (61)

Prior chemotherapy

Fluoropyrimidine plus platinum 38 (93)

Fluoropyrimidine plus other 1 (2)

Platinum plus other 2 (5)

HER2 status 24 (100)

Negative 18 (75.0)

Not done 1 (4.2)

Positive 3 (12.5)

Unknown 2 (8.3)

NSCLC

Adenocarcinoma 22 (81)

Squamous cell carcinoma 4 (15)

NSCLC-NOS 1 (4)

Prior chemotherapy

Pemetrexed 20 (74)

Carboplatin 14 (52)

Cisplatin 15 (56)

Taxane 14 (52)

EGFR mutation status 27 (100)

Negative 23 (85.2)

Not done 3 (11.1)

Positive 1 (3.7)

ALK mutation status 27 (100)

Negative 22 (81.5)

Not done 4 (14.8)

Unknown 1 (3.7)

UC

Bladder as primary site 16 (67)

Haemoglobin level <10 g/dL 3 (13)

Liver metastases 7 (29)

Visceral metastases 24 (100)

Bellmunt risk factors

0 2 (8.3)

1 9 (38)

2 9 (38)

3 4 (17)

Prior chemotherapy

Gemcitabine 24 (100)

Cisplatin 19 (79)

Carboplatin 10 (42)

Data are n (%) unless otherwise indicated. ECOG=Eastern Cooperative Oncology Group. G-ALK=anaplastic lymphoma kinase; EGFR=epidermal growth factor receptor; GEJ=gastric or gastro-oesophageal junction. HER2=human epidermal growth factor receptor 2; NOS=not otherwise specified. NSCLC=non-small-cell lung cancer. PD-L1=programmed death ligand 1. TPS=tumour proportion score. UC=urothelial carcinoma.*Although the protocol limited eligibility to those with ECOG PS 0 to 1, two patients whose ECOG PS changed from 1 to 2 after screening were included.#A detailed summary of prior anticancer therapies is included in table S1.

Table 2: Treatment-related adverse events (n)

  G-GEJn=41

NSCLCn=27

UCn=24

Grade

1-2

Grade

3

Grade

4

Grade

5

Grade

1-2

Grade

3

Grade

4

Grade

5

Grade

1-2

Grade

3

Grade

4

Grade

5

Fatigue* 16 (39) 2 (5) - - 8 (30) 1 (4) - - 6 (25) - - -

Hypothyroidism 4 (10) - - - 7 (26) - - - 2 (8) - - -

Hypertension 2 (5) 3 (7) - - 6 (22) 2 (7) - - 2 (8) 1 (4) - -

Rash* 8 (20) - - - 1 (4) - - - - - - -

Diarrhoea 1 (2) 1 (2) - - 5 (19) - - - 1 (4) 1 (4) - -

Nausea 3 (7) - - - 4 (15) - - - 4 (17) - - -

Pyrexia 2 (5) - - - - - - - 4 (17) - - -

Epistaxis 3 (7) - - - 4 (15) - - - - - - -

Stomatitis 1 (2) 1 (2) - - 4 (15) - - - - - - -

Proteinuria 1 (2) - - - 2 (7) 1 (4) - - 3 (13) - - -

ALT increased 1 (2) - - - 2 (7) - - - 3 (13) - - -

AST increased 2 (5) - - - 2 (7) - - - 3 (13) - - -

Infusion-related reaction 5 (12) 1 (2) - - 1 (4) 1 (4) - - 1 (4) - - -

Decreased appetite 5 (12) - - - 3 (11) 1 (4) - - 1 (4) - - -

Pruritus 5 (12) - - - 2 (7) - - - 1 (4) - - -

Vomiting 1 (2) - - - 3 (11) - - - 2 (8) - - -

Arthralgia - - - - 3 (11) - - - 1 (4) - - -

Headache 1 (2) - - - 3 (11) - - - 2 (8) - - -

Anaemia 1 (2) - 1 (2) - 3 (11) - - - 1 (4) - - -

Dysphonia 1 (2) - - - 3 (11) - - - - - - -

Metrorrhagia† 1 (10) - - - - - - - - - - -

Colitis - 3 (7) - - - - - - - 2 (8) - -

Abdominal pain* - 3 (7) - - 2 (7) - - - 1 (4) - - -

Myocardial infarction - - - - - 2 (7) - - - - - -

Pneumonitis - 1 (2) - - - - - - 1 (4) - - -

Pulmonary embolism 1 (2) 1 (2) - - - - - - - 1 (4) - -

Hyponatremia - - - - - 1 (4) - - - 1 (4) - -

Hypoalbuminaemia - - - - - 1 (4) - - - - - -

Hypophysitis - - - - 1 (4) 1 (4) - - - - - -

Hypokalaemia - - - - - - 1 (4) - - - - -

Delirium - - - - - 1 (4) - - - - - -

Respiratory failure - - - - - 1 (4) - - - - - -

Adrenal insufficiency - - - - - 1 (4) - - 1 (4) - - -

Dehydration - - - - - 1 (4) - - - - - -

Cholestasis - - 1 (2) - - - - - - - - -

PCP Pneumonia - - 1 (2) - - - - - - - - -

Pulmonary sepsis - - - 1 (2) - - - - - - - -

Interstitial lung disease - 1 (2) - - - - - - - - - -

Dyspnoea 1 (2) 1 (2) - - - - - - - - - -

Acute kidney injury - 1 (2) - - - - - - - - - -

Peritonitis - 1 (2) - - - - - - - - - -

Jaundice - 1 (2) - - - - - - - - - -

Hepatitis - 1 (2) - - - - - - - - - -

Data are number (%). The table shows treatment-related adverse events occurring at Grade 1-2 in at least 10% of treated patients per tumour type, or at grades 3-5 in one or more patients according to preferred term or *consolidated categories. †Denominator adjusted because gender specific event for females (N=10).AST=aspartate aminotransferase. ALT=alanine aminotransferase. G-GEJ=gastric or gastro-oesophageal junction adenocarcinoma. NSCLC=non-small-cell lung cancer. PCP=pneumocystis jirovecii pneumonia. UC=urothelial carcinoma.

Table 3: Confirmed efficacy results per RECIST  G-GEJ

n=41NSCLC

n=27UC

n=24Median follow-up duration, months (IQR) 32.2 (24.4-33.7) 33.4 (32.8-33.6) 30.7 (11.3-30.9)

Best overall response, No. (%)

Complete response 0 1 (4) 0

Partial response 3 (7) 7 (26) 3 (13)

Stable disease 18 (44) 15 (56) 9 (38)

Progressive disease 13 (32) 3 (11) 11 (46)

Not evaluable 7 (17) 1 (4) 1 (4)

Objective response, % (95% CI) 7 (1.5-19.9) 30 (13.8-50.2) 13 (2.7-32.4)

Disease control rate, % (95% CI) 51 (35.1-67.1) 85 (66.3-95.8) 50 (29.1-70.9)

Median time to response, months (95% CI) 1.4 (1.4-4.1) 2.1 (1.3-3.0) 2.8 (1.3-5.5)

Median duration of response, months

(95% CI)6.7 (4.4-17.5) NR (11.1-NR) 8.3 (4.6-16.8)

Median duration of stable disease, months

(95% CI)5.0 (4.0-8.5) 6.9 (2.8-9.7) 2.8 (1.9-13.1)

Progression-free survival

Events, n (%) 38 (93) 17 (63) 21 (88)

Median, months (95% CI) 2.5 (1.5-4.2) 9.7 (4.6-27.6) 1.9 (1.2-2.8)

3-month rate, % (95% CI) 44 (28.0-58.4) 77 (55.7-88.9) 28 (11.5-47.2)

6-month rate, % (95% CI) 26 (13.4-40.0) 65 (43.0-79.9) 23 (8.5-42.2)

12-month rate, % (95% CI) 12 (3.9-24.3) 43 (23.4-61.5) 11.6 (2.2-29.9)

Overall survival

Deaths, n (%) 33 (81) 15 (56) 18 (75)

Median, months (95% CI) 5.9 (4.4-10.6) 26.2 (11.8-NR) 6.4 (2.5-18.7)

6-month rate, % (95% CI) 49 (32.9-63.6) 85 (64.5-94.0) 52 (30.5-70.2)

12-month rate, % (95% CI) 31 (16.9-45.3) 68 (45.8-82.6) 38 (18.6-57.3)

18-month rate, % (95% CI) 22 (10.6-36.5) 59 (37.6-75.8) 32 (13.4-51.8)

CI=confidence interval. G-GEJ=gastric or gastro-oesophageal junction. IQR=interquartile range. NR=not reached. NSCLC=non-small-cell lung cancer. RECIST, Response Evaluation Criteria in Solid Tumors. UC=urothelial carcinoma.

Figure 1: Disposition of patients. G-GEJ=gastric or gastro-oesophageal junction. NSCLC=non-small-cell cancer

Figure 2: G-GEJ (Cohorts A/B)

(A) Best percentage change of targeted lesions from baseline versus treatment duration in patients

with G-GEJ. Patients (x-axis) were ordered by percentage change of targeted lesions from baseline

(y-axis) and colour-coded for best response according to PD-L1 expression. Treatment duration (z-

axis) is shown as on or off study at time of data cut-off. (B) Kaplan-Meier plot for progression-free

survival. (C) Kaplan-Meier plot for overall survival. G-GEJ=gastric or gastro-oesophageal junction.

PD-L1=programmed death ligand 1.

Figure 3: NSCLC (Cohort C)

(A) Best percentage change of targeted lesions from baseline versus treatment duration in patients

with NSCLC. Patients (x-axis) were ordered by percentage change of targeted lesions from baseline

(y-axis) and colour-coded for best response according to PD-L1 expression. Treatment duration (z-

axis) is shown as on or off study at time of data cut-off. (B) Kaplan-Meier plot for progression-free

survival. (C) Kaplan-Meier plot for overall survival. NSCLC=non-small-cell lung cancer. PD-

L1=programmed death ligand 1.

Figure 4: UC (Cohort D)

(A) Best percentage change of targeted lesions from baseline versus treatment duration in patients

with UC. Patients (x-axis) were ordered by percentage change of targeted lesions from baseline (y-

axis) and colour-coded for best response according to PD-L1 expression. Treatment duration (z-axis)

is shown as on or off study at time of data cut-off. (B) Kaplan-Meier plot for progression-free survival.

(C) Kaplan-Meier plot for overall survival. PD-L1, programmed death ligand 1. UC=urothelial

carcinoma.